Semantic Segmentation on DELIVER (Condition Breakdown)
70.49mIoU (Mean)SGMA
Evaluation results
| Method | Links | |||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SGMABackbone=Pvt-v2-b22026.03 | 70.49 | — | — | — | — | — | — | — | — | — | — | — | 71.07 | 73.45 | 47.18 | 45.22 | 80.67 | 71.54 | 71.93 | 76.08 | 74.96 | 55.55 | 79.53 | 80.73 | 72.22 | 76.41 | 80.76 | 80.76 | 45.22 | |
| SGMABackbone=ResNet-502026.03 | 68.58 | — | — | — | — | — | — | — | — | — | — | — | 67.69 | 71.46 | 43.01 | 43.47 | 78 | 70.84 | 72.43 | 73.38 | 71.75 | 52.52 | 78.18 | 78.28 | 71.25 | 77.8 | 78.63 | 78.63 | 43.01 | |
| CMNeXtModality=RGB-D-E-L, #Params(M)=58.73, GFLOPS=65.422023.03 | 66.3 | 68.7 | 65.67 | 62.46 | 67.5 | 66.57 | 62.91 | 64.59 | 60 | 65.92 | 65.48 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMNeXtModality=RGB-D-E, #Params(M)=58.72, GFLOPS=64.192023.03 | 64.44 | 68.28 | 63.28 | 62.64 | 63.01 | 66.06 | 62.58 | 64.44 | 58.73 | 65.37 | 65.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| IMLTBackbone=ResNet-502026.03 | 63.85 | — | — | — | — | — | — | — | — | — | — | — | 69.66 | 67.85 | 27.23 | 31.46 | 75.35 | 69.85 | 70.23 | 70.88 | 70.32 | 33.96 | 76.46 | 76.72 | 68.46 | 72.34 | 77.04 | 77.04 | 27.23 | |
| CMNeXtModality=RGB-D, #Params(M)=58.69, GFLOPS=62.942023.03 | 63.58 | 67.21 | 62.79 | 61.64 | 62.95 | 65.26 | 61 | 64.64 | 58.71 | 64.32 | 63.35 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMXModality=RGB-D, #Params(M)=66.57, GFLOPS=65.682023.03 | 62.66 | 63.7 | 62.77 | 60.74 | 62.37 | 63.14 | 59.5 | 60.14 | 55.84 | 62.65 | 63.26 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| IMLTBackbone=Pvt-v2-b22026.03 | 62.33 | — | — | — | — | — | — | — | — | — | — | — | 55.54 | 69.19 | 29.82 | 32.87 | 76.46 | 57.93 | 58.61 | 73.95 | 73.43 | 37.25 | 77.47 | 77.72 | 61.23 | 75.35 | 78.08 | 78.08 | 29.82 | |
| SegFormerModality=RGB, #Params(M)=25.79, GFLOPS=38.932023.03 | 57.2 | 59.99 | 57.3 | 50.45 | 58.69 | 60.21 | 57.28 | 56.64 | 37.44 | 57.17 | 59.12 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SGMABackbones=Pvt-v2-b22026.03 | 55.49 | — | — | — | — | — | — | — | — | — | — | — | 55.12 | 58.04 | 30.87 | 29.22 | 67.6 | 55.69 | 56.16 | 61.4 | 59.95 | 38.46 | 66.01 | 67.69 | 56.52 | 61.82 | 67.73 | 67.73 | 29.22 | |
| MAGICBackbone=ResNet-502026.03 | 54.76 | — | — | — | — | — | — | — | — | — | — | — | 62.78 | 59.06 | 2.96 | 12.35 | 73.76 | 62.65 | 62.8 | 62.14 | 62.21 | 13.14 | 74.27 | 74.1 | 61.74 | 63.32 | 74.12 | 74.27 | 2.96 | |
| MAGICBackbone=Pvt-v2-b22026.03 | 54.59 | — | — | — | — | — | — | — | — | — | — | — | 46.11 | 71.63 | 7.21 | 8.83 | 78.06 | 46.29 | 49.43 | 72.38 | 71.85 | 10.77 | 77.8 | 77.68 | 51.02 | 71.72 | 78.08 | 78.08 | 7.21 | |
| SGMABackbones=ResNet-502026.03 | 53.28 | — | — | — | — | — | — | — | — | — | — | — | 51.16 | 55.59 | 27.4 | 27.77 | 63.93 | 54.85 | 56.78 | 57.95 | 55.95 | 35.61 | 64.17 | 64.31 | 55.34 | 63.66 | 64.79 | 64.79 | 27.4 | |
| HRFuserModality=RGB-D-E-L, #Params(M)=31.61, GFLOPS=235.02023.03 | 52.97 | 56.2 | 52.39 | 49.85 | 52.53 | 54.02 | 49.44 | 46.31 | 46.92 | 53.94 | 52.72 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HRFuserModality=RGB-D, #Params(M)=30.46, GFLOPS=223.02023.03 | 51.88 | 54.8 | 51.48 | 49.51 | 51.55 | 52.12 | 50.92 | 41.51 | 44 | 54.1 | 52.52 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HRFuserModality=RGB-D-E, #Params(M)=31.04, GFLOPS=229.02023.03 | 51.83 | 54.04 | 50.83 | 50.88 | 51.13 | 52.61 | 49.32 | 41.75 | 47.89 | 54.65 | 52.33 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TokenFusionModality=RGB-D, #Params(M)=26.01, GFLOPS=54.962023.03 | 49.86 | 50.92 | 52.02 | 43.37 | 50.7 | 52.21 | 49.22 | 46.22 | 36.39 | 49.58 | 49.17 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| IMLTBackbones=ResNet-502026.03 | 48.83 | — | — | — | — | — | — | — | — | — | — | — | 53.45 | 51.34 | 15.76 | 18.67 | 60.45 | 53.67 | 54.12 | 54.89 | 54.23 | 20.45 | 61.89 | 62.23 | 52.05 | 56.67 | 62.65 | 62.65 | 15.76 | |
| HRFuserModality=RGB, #Params(M)=29.89, GFLOPS=217.52023.03 | 47.95 | 49.26 | 48.64 | 42.57 | 50.61 | 50.47 | 48.33 | 35.13 | 26.86 | 49.06 | 49.88 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| M3LBackbone=Pvt-v2-b22026.03 | 47.64 | — | — | — | — | — | — | — | — | — | — | — | 31.47 | 54.49 | 4.15 | 2.63 | 74.83 | 39.01 | 42.94 | 58.72 | 50.88 | 8.52 | 77.22 | 77.4 | 45.63 | 69.85 | 76.89 | 77.4 | 2.63 | |
| IMLTBackbones=Pvt-v2-b22026.03 | 47.17 | — | — | — | — | — | — | — | — | — | — | — | 38.45 | 52.89 | 17.52 | 19.67 | 61.89 | 40.78 | 41.45 | 58.67 | 58.01 | 22.89 | 63.23 | 63.56 | 44.12 | 60.45 | 64.04 | 64.04 | 17.52 | |
| M3LBackbone=ResNet-502026.03 | 42.75 | — | — | — | — | — | — | — | — | — | — | — | 25.42 | 44.16 | 2.43 | 1.71 | 71.52 | 31.82 | 35.32 | 49.37 | 41.95 | 6.67 | 75.35 | 75.61 | 38.28 | 65.07 | 76.63 | 76.63 | 1.71 | |
| MAGICBackbones=Pvt-v2-b22026.03 | 41.39 | — | — | — | — | — | — | — | — | — | — | — | 29.96 | 55.8 | 3.74 | 4.62 | 64.01 | 30.12 | 32.83 | 56.71 | 56.07 | 5.69 | 63.66 | 63.5 | 34.25 | 55.91 | 64.04 | 64.04 | 3.74 | |
| MAGICBackbones=ResNet-502026.03 | 40.71 | — | — | — | — | — | — | — | — | — | — | — | 45.75 | 41.91 | 1.5 | 6.58 | 58.43 | 45.61 | 45.77 | 45.07 | 45.15 | 7.03 | 59.07 | 58.85 | 44.65 | 46.33 | 58.88 | 59.07 | 1.5 | |
| M3LBackbones=Pvt-v2-b22026.03 | 34.85 | — | — | — | — | — | — | — | — | — | — | — | 18.67 | 37.45 | 2.12 | 1.33 | 59.78 | 24.23 | 27.34 | 41.56 | 34.12 | 4.45 | 62.89 | 63.13 | 29.56 | 53.67 | 62.45 | 63.13 | 1.33 | |
| MuSSBackbone=Pvt-v2-b22026.03 | 34.13 | — | — | — | — | — | — | — | — | — | — | — | 7.51 | 1.39 | 2.27 | 1.07 | 66.8 | 23.2 | 30.82 | 35.34 | 7.71 | 5.83 | 79.69 | 79.96 | 27.23 | 63.28 | 79.82 | 79.96 | 1.07 | |
| MuSSBackbone=ResNet-502026.03 | 32.45 | — | — | — | — | — | — | — | — | — | — | — | 1.8 | 0.68 | 1.13 | 0.78 | 63.88 | 17.88 | 24.15 | 5.03 | 3.36 | 5.86 | 77.8 | 77.7 | 50.8 | 77.83 | 78.03 | 78.03 | 0.68 | |
| M3LBackbones=ResNet-502026.03 | 30.6 | — | — | — | — | — | — | — | — | — | — | — | 14.56 | 28.34 | 1.23 | 0.86 | 55.67 | 18.92 | 21.45 | 32.78 | 26.54 | 3.45 | 60.45 | 60.78 | 23.67 | 48.23 | 62.11 | 62.11 | 0.86 | |
| MuSSBackbones=Pvt-v2-b22026.03 | 25.17 | — | — | — | — | — | — | — | — | — | — | — | 3.9 | 0.7 | 1.15 | 0.54 | 50.15 | 13.12 | 18.22 | 21.46 | 4.01 | 3 | 66.23 | 66.61 | 15.76 | 46.29 | 66.41 | 66.61 | 0.54 | |
| MuSSBackbones=ResNet-502026.03 | 24.6 | — | — | — | — | — | — | — | — | — | — | — | 0.91 | 0.34 | 0.57 | 0.39 | 46.93 | 9.82 | 13.73 | 2.58 | 1.71 | 3.02 | 63.67 | 63.53 | 34.05 | 63.7 | 63.98 | 63.98 | 0.34 | |
| CMNextBackbone=MiT-B2, Params=58.7M, Flops=62.9G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 63.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMNextBackbone=MiT-B4, Params=116.6M, Flops=112.0G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 66.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMXBackbone=MiT-B2, Params=66.6M, Flops=65.7G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 62.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DFormerv2-BBackbone=DFormerv2-B, Params=53.9M, Flops=60.8G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 65.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DFormerv2-LBackbone=DFormerv2-L, Params=95.5M, Flops=114.5G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 67.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DFormerv2-SBackbone=DFormerv2-S, Params=26.7M, Flops=28.9G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 63.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GeminiFusionBackbone=MiT-B5, Params=137.2M, Flops=218.4G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 66.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HRFuserBackbone=HRFormer-T, Params=30.5M, Flops=223.0G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 51.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TokenFusionBackbone=MiT-B2, Params=26.0M, Flops=55.0G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 60.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TokenFusionBackbone=MiT-B5, Params=83.3M, Flops=144.7G, Resolution=512x5122025.04 | — | — | — | — | — | — | — | — | — | — | — | 63.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |